In the continuous casting production process, defect detection of surface corrugation is a key process to ensure the quality of the hot rolled sheet products in the subsequent process. Currently, the manual detection approaches used in most steel productions inevitably causes an ineffectiveness and inefficiency. Accordingly, an effective detection method is urgently needed to identify the defections of steel plate corrugations. Aiming at this problem, this paper designs a new defect detection method based on image cropping and region growing algorithm, which uses image detection technology to detect steel plate ripple and to determine whether there is a defect. First, based on a detailed analysis of the steel plate surface defects, we combine the image cropping with improved region growing algorithm to deal with the defect image. Especially, the portion with no defects is cut from in its image by image cropping algorithm, and then the rest are detected by the region growing algorithm. Finally, this algorithm is validated on steel plate data sets. Experimental results show that the proposed algorithm performs powerfully, obtaining a higher recognition rate. That means this algorithm is an effective intelligent detection tool to replacing traditional manual recognition in some ways.
展开▼
